Premium
Model predictive control technique combined with iterative learning for batch processes
Author(s) -
Lee Kwang S.,
Chin InShik,
Lee Hyuk J.,
Lee Jay H.
Publication year - 1999
Publication title -
aiche journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.958
H-Index - 167
eISSN - 1547-5905
pISSN - 0001-1541
DOI - 10.1002/aic.690451016
Subject(s) - iterative learning control , model predictive control , batch reactor , batch processing , nonlinear system , control theory (sociology) , process (computing) , trajectory , computer science , tracking (education) , feature (linguistics) , process control , control engineering , control (management) , engineering , artificial intelligence , chemistry , psychology , pedagogy , biochemistry , physics , operating system , linguistics , philosophy , programming language , catalysis , quantum mechanics , astronomy
A novel model predictive control technique geared specifically toward batch process applications is demonstrated in an experimental batch reactor system for temperature tracking control. The technique, called Batch‐MPC (BMPC), is based on a time‐varying linear system model (representing a nonlinear system along a fixed trajectory) and utilizes not only the incoming measurements from the ongoing batch, but also the information stored from the past batches. This particular feature is shown to be essential for achieving effective tracking control performance despite model errors and disturbances. In a series of experiments performed on a bench‐scale batch reactor system, the technique was found to deliver satisfactory tracking performance, as expected, overcoming a large amount of model uncertainty and various process disturbances.